DATA PROCESSING METHOD AND RELATED DEVICE

WIPO Patent Application WO/2024/082891

A1

Disclosed in embodiments of the present application is a data processing method. The method is applied to a text recognition/character recognition scenario. The method comprises: acquiring input data, the input data being image data or audio data, and acquiring a second modal feature according to a first modal feature of the input data, the first modal feature being a visual feature of the image data or an audio feature of the audio data, and the second modal feature being a character feature; and fusing the first modal feature and the second modal feature to obtain a target feature. Information of different modal data can be efficiently fused, to cause the obtained target feature to have characteristics of multi-modal data, so that the expression capability of the target feature is improved, thereby increasing the precision of a first recognition result obtained according to the target feature. Moreover, compared with a method for determining a recognition result only according to a corrected second modal feature, the re-introduction of the first modal feature before correction can prevent over-correction of the second modal feature.

FU YIFEI (CN)
HU HAILIN (CN)
ZHU MINGJIAN (CN)
CHEN XINGHAO (CN)
WANG YUNHE (CN)

PCT/CN2023/119082

April 25, 2024

September 15, 2023

Click for automatic bibliography generation  

HUAWEI TECH CO LTD (CN)

G06V20/62

CN116434752A	2023-07-14
CN112257426A	2021-01-22
CN111738251A	2020-10-02
CN115116444A	2022-09-27
CN112687296A	2021-04-20
CN113822340A	2021-12-21
JP2019074807A	2019-05-16

SHENPAT INTELLECTUAL PROPERTY AGENCY (CN)

View/Download PDF      PDF Help

Previous Patent: SCHOOL BADGE HAVING MOTION POSITIONING FUNCTION

Next Patent: SCHEDULING METHOD FOR UWB AND RELATED PRODUCT